DATE:
AUTHOR:
The Launchable Product Team
CLI Predictive Test Selection

Use 'flavors' to run the best tests for an environment

DATE:
AUTHOR: The Launchable Product Team

Lots of teams run the same tests across several different environments. For example, a UI test suite might be run in several browsers in parallel. Or perhaps you need to build a slightly different version of a mobile app for different locales and need to run the same tests across all of them.

In these scenarios, a test result is not just a test result: it is the combination of the test and the environment that it was run in.

Launchable now supports these scenarios with a new concept called flavors.

When you submit test results using launchable record tests, you can now submit additional metadata in the form of key-value pairs using the ---flavor option.

For example:

// after running tests in Chrome
launchable record tests --build [BUILD NAME] --flavor browser=chrome cypress ./report-chrome.xml
// after running tests in Firefox
launchable record tests --build [BUILD NAME] --flavor browser=firefox cypress ./report-chrome.xml

And so on. (You can submit multiple key-value pairs, too.)

Later, when you want to request a subset of tests, you can include the same key-value pairs to get a subset of tests specifically selected for that flavor.

For example:

// before running tests in Chrome
find ./cypress/integration | launchable subset --build [BUILD NAME] --confidence 90% --flavor browser=chrome cypress > subset-chrome.txt
// run the Chrome subset
cypress run --spec "$(cat subset-chrome.txt)" --reporter junit --reporter-options "mochaFile=report/test-output-chrome-[hash].xml"
...
// before running tests in Firefox
find ./cypress/integration | launchable subset --build [BUILD NAME] --confidence 90% --flavor browser=firefox cypress > subset-chrome.txt
// run the Chrome subset
cypress run --spec "$(cat subset-chrome.txt)" --reporter junit --reporter-options "mochaFile=report/test-output-firefox-[hash].xml"

This feature lets you select the right tests to run based on the changes being tested and the environment they are being run in.

Powered by LaunchNotes